SBIR Phase I: Software Defined Networking with Partially Ordered Multipath Routing

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project will be a new network architecture to improve the internet. Its original connectionless, packet-switched architecture was a departure from the traditional circuit-switched model and involved significant risks. To manage these risks, the original design focused on a simple network with smart endpoints. The result was scalable, robust and supported a wide range of communications technologies; however, it is limited in its ability to meet the performance, security, and policy needs of many modern network applications. Overprovisioning and expensive, complex, and fragile add-on technology are required to meet the needs of these new applications and these solutions may not scale This project will provide an enhanced, smart network that is scalable and robust, supporting performance and policy control while making efficient use of network resources. This will improve the next generation of internet applications. This Small Business Innovation Research (SBIR) Phase I project explores the translation of networking with constraints. In this architecture, performance requirements are expressed as performance constraints, and policy requirements for security, multi-tenancy, and traffic differentiation as Boolean constraints on the resources used for an application. Routing with constraints computes the best set of paths per destination that provide the full range of performance and policies supported by the network, allowing traffic to be sent over paths that meet the needs of applications, and distributing the load more evenly over the network (simulations show a 10x increase in capacity). This approach is more robust because it is implemented as a part of the routing function, directly responding to network changes, and more efficient because it runs on the "native internet," eliminating overlays. Lastly, it improves security by implementing a default-deny communication model and is easier to configure based on its declarative, "what" configuration model. This project is focused on characterizing and mitigating risks remaining in this architecture and demonstrating the feasibility of this approach through the development of an operational proof-of-concept. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.